Journal Articles
Permanent URI for this collectionhttps://mro.massey.ac.nz/handle/10179/7915
Browse
6 results
Search Results
Item Large Multi-Modal Model Cartographic Map Comprehension for Textual Locality Georeferencing(Schloss Dagstuhl – Leibniz-Zentrum für Informatik, 2025-08-15) Wijegunarathna K; Stock K; Jones CB; Sila-Nowicka K; Moore A; O’Sullivan D; Adams B; Gahegan MMillions of biological sample records collected in the last few centuries archived in natural history collections are un-georeferenced. Georeferencing complex locality descriptions associated with these collection samples is a highly labour-intensive task collection agencies struggle with. None of the existing automated methods exploit maps that are an essential tool for georeferencing complex relations. We present preliminary experiments and results of a novel method that exploits multimodal capabilities of recent Large Multi-Modal Models (LMM). This method enables the model to visually contextualize spatial relations it reads in the locality description. We use a grid-based approach to adapt these auto-regressive models for this task in a zero-shot setting. Our experiments conducted on a small manually annotated dataset show impressive results for our approach (∼1 km Average distance error) compared to uni-modal georeferencing with Large Language Models and existing georeferencing tools. The paper also discusses the findings of the experiments in light of an LMM's ability to comprehend fine-grained maps. Motivated by these results, a practical framework is proposed to integrate this method into a georeferencing workflow.Item Predicting Distance and Direction from Text Locality Descriptions for Biological Specimen Collections(Schloss Dagstuhl – Leibniz-Zentrum für Informatik, 2022-08-22) Liao R; Das PP; Jones CB; Aflaki N; Stock K; Ishikawa T; Fabrikant SI; Winter SA considerable proportion of records that describe biological specimens (flora, soil, invertebrates), and especially those that were collected decades ago, are not attached to corresponding geographical coordinates, but rather have their location described only through textual descriptions (e.g. North Canterbury, Selwyn River near bridge on Springston-Leeston Rd). Without geographical coordinates, millions of records stored in museum collections around the world cannot be mapped. We present a method for predicting the distance and direction associated with human language location descriptions which focuses on the interpretation of geospatial prepositions and the way in which they modify the location represented by an associated reference place name (e.g. near the Manawatu River). We study eight distance-oriented prepositions and eight direction-oriented prepositions and use machine learning regression to predict distance or direction, relative to the reference place name, from a collection of training data. The results show that, compared with a simple baseline, our model improved distance predictions by up to 60% and direction predictions by up to 31%.Item The Koja Web Mapping Application for Context-sensitive Natural Language Spatial Querying(CEUR Team, 2023-01-01) Aflaki N; Stock K; Jones CB; Guesgen H; Fukuzawa Y; Morley J; Hu X; Hu Y; Resch B; Kersten J; Stock KThe locations of objects are often described in natural language relative to some other object using vague and context-sensitive spatial relation terms (e.g. theatre near Trafalgar Square). Koja is a web map application that predicts the distance between a location and reference object based on the spatial relation term specified by the user and language and contextual features. That distance is used to retrieve objects of the specified type within a range of the distance. They are displayed through a map interface to make the process more intuitive and user-friendly.Item AI-based discovery of habitats from museum collections(Cell Press, 2024-04-02) Jones CB; Stock K; Perkins SEMuseum collection records are a source of historic data for species occurrence, but little attention is paid to the associated descriptions of habitat at the sample locations. We propose that artificial intelligence methods have potential to use these descriptions for reconstructing past habitat, to address ecological and evolutionary questions.Item Detecting geospatial location descriptions in natural language text(Taylor and Francis Group, 2022) Stock K; Jones CB; Russell S; Radke M; Das P; Aflaki NReferences to geographic locations are common in text data sources including social media and web pages. They take different forms from simple place names to relative expressions that describe location through a spatial relationship to a reference object (e.g. the house beside the Waikato River). Often complex, multi-word phrases are employed (e.g. the road and railway cross at right angles; the road in line with the canal) where spatial relationships are communicated with various parts of speech including prepositions, verbs, adverbs and adjectives. We address the problem of automatically detecting relative geospatial location descriptions, which we define as those that include spatial relation terms referencing geographic objects, and distinguishing them from non-geographical descriptions of location (e.g. the book on the table). We experiment with several methods for automated classification of text expressions, using features for machine learning that include bag of words that detect distinctive words, word embeddings that encode meanings of words and manually identified language patterns that characterise geospatial expressions. Using three data sets created for this study, we find that ensemble and meta-classifier approaches, that variously combine predictions from several other classifiers with data features, provide the best F-measure of 0.90 for detecting geospatial expressions.Item Multi-source Multimodal Data and Deep Learning for Disaster Response: A Systematic Review(Springer Nature, 2022-01) Algiriyage N; Prasanna R; Stock K; Doyle EEH; Johnston DMechanisms for sharing information in a disaster situation have drastically changed due to new technological innovations throughout the world. The use of social media applications and collaborative technologies for information sharing have become increasingly popular. With these advancements, the amount of data collected increases daily in different modalities, such as text, audio, video, and images. However, to date, practical Disaster Response (DR) activities are mostly depended on textual information, such as situation reports and email content, and the benefit of other media is often not realised. Deep Learning (DL) algorithms have recently demonstrated promising results in extracting knowledge from multiple modalities of data, but the use of DL approaches for DR tasks has thus far mostly been pursued in an academic context. This paper conducts a systematic review of 83 articles to identify the successes, current and future challenges, and opportunities in using DL for DR tasks. Our analysis is centred around the components of learning, a set of aspects that govern the application of Machine learning (ML) for a given problem domain. A flowchart and guidance for future research are developed as an outcome of the analysis to ensure the benefits of DL for DR activities are utilized.